Back to Employee Satisfaction & Pulse Surveys

Test and QA Plan

How the build is validated before go-live: three sequential test phases, edge cases, and production verification.

4 pagesPDF · Quality

FS-DOC-06Quality

Test and QA Plan

Employee Satisfaction and Pulse Surveys

[YourCompany.com] · HR Department · Prepared by FullSpec · [Today's Date]

This document defines the structured test and quality assurance programme for the Employee Satisfaction and Pulse Survey automation. It covers three sequential phases: unit testing of individual agent logic, integration and edge-case testing across agent handoffs and failure scenarios, and a final end-to-end production run with process owner sign-off. The FullSpec team executes Phases 1 and 2 in sandbox environments; the process owner joins the FullSpec team for Phase 3. No phase begins until all test cases in the preceding phase achieve a full pass.

01Testing approach

Testing is structured across three phases that must be completed in strict sequence. Phase 1 validates each agent in isolation against known inputs. Phase 2 confirms that agents hand off correctly to one another and that the system handles edge cases gracefully. Phase 3 runs a full production cycle end to end and records the process owner's formal approval. No phase begins until every test case in the previous phase reaches a passing state and is documented by the FullSpec team.

Phase 1: Unit Testing

Individual agent logic in isolation

Scope

Each agent tested independently: Survey Distribution Agent and Response Analysis Agent. No cross-agent calls during this phase.

Environment

Sandbox only. All tool connections (BambooHR, Typeform, Gmail, Google Sheets, Notion, Slack) use sandbox or test credentials.

Credential rule

Sandbox API keys must be used throughout. No production credentials are loaded into the automation platform during Phase 1.

Pass condition

All test cases T01 through T12 must return the expected output with zero unhandled errors before Phase 2 may begin.

Who runs it

FullSpec team only.

Phase 2: Integration and Edge-Case Testing

Agent handoffs, tool integrations, and failure modes

Scope

End-to-end agent handoffs, tool-to-tool data flows, and all identified edge cases including duplicate detection, missing data handling, and tool unavailability.

Environment

Sandbox for all automated steps. A staging Notion workspace and a test Slack channel are used. No production data is written.

Credential rule

Sandbox credentials remain active. Production credentials must not be substituted at any point during Phase 2.

Pass condition

All integration test cases T13 through T20 and edge-case tests E01 through E07 must pass before Phase 3 may begin.

Who runs it

FullSpec team only.

Phase 3: End-to-End Production Test

Full production run with process owner review and approval

Scope

One complete survey cycle run against live production systems: BambooHR production API, live Gmail sending, Typeform production form, Google Sheets production sheet, Notion production workspace, and Slack production channel.

Environment

Production. A dedicated internal test address is used for all email sends. The test survey record and any Notion draft created are voided after completion.

Credential rule

Production credentials are loaded for this phase only. Sandbox credentials are revoked from the platform before Phase 3 begins.

Pass condition

Test cases T21 through T25 must all pass and the process owner must complete T25 (review and approval) with a recorded confirmation logged by FullSpec before go-live is authorised.

Who runs it

FullSpec team, with the process owner (HR Manager) joining for T25.

Test and QA PlanPage 1 of 4

FS-DOC-06Quality

02Phase 1 — unit test cases

Sandbox credentials only. All Phase 1 tests must be executed with sandbox or test API keys for BambooHR, Typeform, Gmail, Google Sheets, Notion, and Slack. Do not load production credentials into the automation platform at any point during this phase. Any test case that touches a live system is invalid and must be re-run under sandbox conditions.

Survey Distribution Agent — Estimated build time: 12 hours — Complexity: Moderate

Scenario

Input

Expected output

Pass criteria

T01

Happy path: fetch active employee list

BambooHR sandbox returns 10 active employees; 2 contractors and 1 leaver present in dataset

Google Sheets log contains exactly 10 rows, contractors and leaver excluded

Row count equals 10; employment_status filter applied correctly; no contractor or inactive record present

T02

Happy path: send personalised invitations

10-employee list from T01; Typeform sandbox survey URL set

10 Gmail sandbox messages sent; each contains employee first name and correct Typeform link; Google Sheets 'invited' column marked TRUE for all 10

Gmail sandbox outbox contains 10 messages; personalisation token resolved for each; Sheets log updated within 30 seconds of send

T03

Response check: all employees responded

Typeform sandbox reports 10 responses matching all 10 invited employee IDs

Decision node evaluates to 'all responded'; no reminder emails sent; flow proceeds to aggregation step

Zero reminder emails in Gmail sandbox outbox; flow log shows 'all_responded: true'

T04

Response check: partial response, reminder fires

Typeform sandbox reports 6 responses; 4 employees have not responded

4 reminder emails sent via Gmail sandbox; each addressed to the correct non-responder; Google Sheets reminder column marked TRUE for those 4; Slack test channel receives response rate alert (60%)

Exactly 4 reminder emails in sandbox outbox; no reminder sent to the 6 who responded; Slack message shows '60% response rate'

T05

Duplicate reminder guard

Reminder trigger fires a second time for the same survey cycle before window closes

System detects reminder_sent flag is TRUE for all non-responders; no second reminder emails sent

Zero additional emails in sandbox outbox on second trigger; log entry reads 'reminder_already_sent: skipped'

T06

BambooHR API returns empty list

BambooHR sandbox returns zero active employees (e.g. misconfigured filter)

Automation halts; error logged with message 'employee_list_empty'; Slack test channel notified; no emails sent

No Gmail messages in sandbox outbox; error record written to Google Sheets error log; Slack alert received within 60 seconds

T07

Gmail send failure on invitation step

Gmail sandbox rate limit simulated; first 3 sends succeed, send 4 returns 429 error

Automation retries send 4 up to 3 times with exponential back-off; on persistent failure writes failed addresses to Google Sheets error log and continues with remaining sends

Sends 1-3 and 5-10 complete; send 4 appears in Sheets error log with status 'send_failed_after_retry'; no silent drop

T08

Typeform response check returns API error

Typeform sandbox returns 503 on response status query

Automation retries up to 3 times; on persistent failure logs error and sends Slack alert; reminder step is skipped until next scheduled check interval

Slack test channel receives 'Typeform unreachable' alert; no reminder emails sent on failed check; retry count recorded in log

Response Analysis Agent — Estimated build time: 16 hours — Complexity: Complex

Scenario

Input

Expected output

Pass criteria

T09

Happy path: aggregate responses into Google Sheets

Typeform sandbox export containing 10 complete responses with eNPS and 5 category scores each

Google Sheets sandbox tab populated with all 10 rows; eNPS calculated correctly; category averages computed for all 5 dimensions

Row count in Sheets equals 10; eNPS formula result matches manual spot-check calculation; no blank cells in score columns

T10

Happy path: Notion report page created

Aggregated scores from T09; prior cycle scores available for trend comparison

Notion sandbox page created with: headline eNPS score, category score table, trend delta vs prior cycle, AI-generated narrative summary, flagged departments below threshold

Notion page exists in sandbox workspace; all five sections present; threshold flag correctly applied to any category average below 6.0 out of 10

T11

Threshold flagging: department below score threshold

One department returns average score of 4.8 (below 6.0 threshold); all others above threshold

Notion page flags that department in red; narrative summary references it explicitly; Slack test channel receives a low-score alert

Flagged department appears in dedicated 'Attention Required' section of Notion page; Slack alert message contains department name and score

T12

Partial Typeform responses excluded correctly

Typeform sandbox export includes 8 complete responses and 2 partially completed (abandoned mid-survey)

Only 8 complete responses written to Google Sheets; partial responses excluded; log records 'partial_responses_excluded: 2'

Sheets row count equals 8; no partial response data present in score columns; exclusion count logged accurately

Test and QA PlanPage 2 of 4

FS-DOC-06Quality

03Phase 2 — integration and edge case tests

Phase 2 validates that the two agents hand off data correctly between each other and that all tool-to-tool connections behave as expected across the full automated flow. All tests remain in the sandbox environment. Integration test IDs continue the T sequence from Phase 1.

Scenario

Input

Expected output

Pass criteria

T13

Survey Distribution Agent to Google Sheets handoff

10 invitations sent and logged by Survey Distribution Agent; Response Analysis Agent trigger fires at window close

Response Analysis Agent reads from the exact Google Sheets tab written by the Distribution Agent; no schema mismatch errors; all 10 rows visible to the analysis step

Analysis Agent reads 10 rows without transformation errors; column headers match expected schema; no missing employee IDs

T14

Typeform to Google Sheets aggregation via Response Analysis Agent

Typeform sandbox webhook fires on survey close with 10 response records

All 10 response records written to the Sheets 'responses' tab within 2 minutes of webhook receipt; eNPS and category scores calculated automatically

Webhook delivery confirmed in Typeform sandbox log; Sheets row count equals 10 within 2-minute SLA; formulas return numeric values for all score columns

T15

Google Sheets to Notion report creation

Scored Sheets tab from T14 used as input; prior cycle data present in Sheets for trend comparison

Notion sandbox page created with all required sections; trend deltas computed from prior cycle; page URL written back to Google Sheets in the 'report_url' column

Notion page accessible via generated URL; all five sections populated; report_url column in Sheets is non-empty

T16

Notion report to Slack and Gmail delivery

Notion sandbox page URL from T15; Slack test channel configured; test Gmail recipient address set

Slack test channel receives message containing Notion page URL and one-line score summary; Gmail sandbox message sent to test recipient with report link and brief summary

Slack message present in test channel within 60 seconds of Notion page creation; Gmail sandbox outbox contains one message to the test address with correct URL

T17

New-hire 30-day trigger via BambooHR

BambooHR sandbox record for a new employee with a hire date exactly 30 days prior to today

Survey Distribution Agent fires independently of the scheduled cycle trigger; new hire receives a standalone survey invitation; invitation logged separately in Google Sheets

Survey invitation in sandbox outbox addressed to the new-hire test email; Google Sheets log contains a row tagged 'trigger_type: new_hire_30_day'

T18

Manual survey close in Typeform triggers Response Analysis Agent

HR manually closes the sandbox Typeform survey before the scheduled window end date

Response Analysis Agent fires within 5 minutes of manual close; aggregation and Notion page creation proceed as per scheduled close

Agent execution log shows trigger source 'typeform_manual_close'; Notion sandbox page created with correct timestamp

T19

End-to-end flow timing: invitation to report

10-employee sandbox run with 8 responding before reminder and 2 responding after reminder

Full flow completes: invitation send, response check, reminder to 2 non-responders, aggregation after close, Notion page creation, Slack and Gmail delivery

All steps complete without manual intervention; total automated processing time from trigger to Slack delivery is under 10 minutes excluding the survey window itself

T20

Google Sheets log accuracy across full cycle

Full sandbox cycle from T19

Google Sheets log contains: 10 invited rows, 8 responded before reminder, 2 reminder flags, 2 responded after reminder, 1 report URL, 0 error rows

Each column in the Sheets log matches the expected state; no duplicate rows; error log tab is empty

Edge-case tests are numbered separately with an E prefix. These scenarios cover failure conditions, data anomalies, and human timeout situations that the automation must handle gracefully.

Scenario

Input

Expected output

Pass criteria

E01

Duplicate employee record in BambooHR export

BambooHR sandbox returns the same employee ID twice in the active list

Deduplication logic removes the duplicate; exactly one invitation sent to that employee; Sheets log shows one row for that employee ID with a 'deduplicated: true' flag

Single invitation in sandbox outbox for the duplicated employee; Sheets log contains one row per employee ID; deduplication flag present

E02

Employee record missing email address in BambooHR

BambooHR sandbox returns one employee record with a null email field

That employee is skipped for invitation; their record is written to the Sheets error log with status 'missing_email'; Slack test channel notified

No invitation sent to a null address; error log row present with employee ID and 'missing_email' status; Slack alert received

E03

Typeform returns zero responses at window close

Survey window closes with zero Typeform responses submitted

Response Analysis Agent fires but detects empty response set; Notion page is not created; Slack alert sent with message 'zero responses received, report not generated'; Google Sheets log records 'responses: 0'

No Notion page created; Slack alert present in test channel; Sheets log reflects zero response count

E04

Google Sheets write failure during invitation logging

Google Sheets API returns 503 during invitation log write after 3 of 10 invitations are sent

Automation retries the Sheets write up to 3 times; on persistent failure continues sending remaining invitations and queues a retry for the Sheets log; Slack alert sent noting partial log failure

Remaining invitations sent; Sheets write retried and completed on subsequent attempt; Slack alert confirms partial log failure and recovery

E05

Notion API unavailable during report creation

Notion sandbox returns 503 during page creation step

Automation retries Notion write up to 3 times with exponential back-off; on persistent failure writes the report data as a structured JSON payload to Google Sheets fallback tab and sends Slack alert

Notion page creation fails gracefully; fallback tab in Sheets contains report payload; Slack alert received with message 'Notion unavailable, report stored in Sheets fallback'

E06

Slack notification fails at report delivery step

Slack sandbox webhook returns 400 during the report delivery notification

Automation logs the Slack failure; Gmail delivery to leadership address continues and succeeds; error logged in Sheets; no silent failure

Gmail delivery completes; Sheets error log records Slack failure with status code; process does not halt on Slack error alone

E07

HR does not review Notion report within 48 hours of creation

Notion page created and HR review step reached; no HR action recorded within 48 hours

Automation sends a follow-up Slack message to the HR Slack channel and a Gmail reminder to the HR Manager address reminding them that the report is pending review before leadership distribution

Slack reminder and Gmail reminder both appear in sandbox outputs at the 48-hour mark; no automatic distribution to leadership until HR review is confirmed

Test and QA PlanPage 3 of 4

FS-DOC-06Quality

04Phase 3 — end-to-end production test

Important before running Phase 3: Use an internal test email address (e.g. a shared HR inbox or a FullSpec test alias) for all invitation and reminder sends during this run. Do not send to real employees. The Notion report draft created during Phase 3 must be deleted or clearly marked as a test record immediately after the process owner completes their review. Any Google Sheets rows written during Phase 3 must be removed from the production sheet before the first live survey cycle runs. Confirm these steps with the FullSpec team before proceeding.

Phase 3 runs a single complete survey cycle against live production systems using a small internal test group. Every step is traced from trigger to Slack and Gmail delivery. The FullSpec team monitors execution in real time and records each result. The process owner joins at T25 to review the Notion report and record their formal approval, which is logged in the FullSpec confirmation record and authorises go-live.

Scenario

What is verified

Expected result

Pass criteria

T21

Production trigger and BambooHR employee fetch

Scheduled trigger fires against production BambooHR API; active employee filter applied to production dataset

Production employee list retrieved with correct active-only filter; row count matches known headcount in BambooHR admin; list written to production Google Sheets log

Production Sheets log row count matches BambooHR admin active headcount; no contractor or inactive record present; trigger timestamp recorded

T22

Production invitation send and Typeform link delivery

Personalised invitations sent via production Gmail to the internal test address only; Typeform production survey link included

Test address receives invitation email with correct first name token resolved and working Typeform link; Google Sheets 'invited' column marked TRUE; no emails sent to any real employee address

One email in production Gmail sent folder addressed to test address; Typeform link opens the correct live survey; Sheets log updated; zero emails to non-test addresses

T23

Production response aggregation and scoring

Test responder submits the Typeform production survey; survey window manually closed by FullSpec team after submission

Response written to production Google Sheets responses tab; eNPS and category scores calculated; prior cycle trend delta computed; no data corruption or formula errors

Sheets row count equals 1 (test submission); all score columns populated with numeric values; eNPS calculation matches manual spot-check; trend delta column non-null

T24

Production Notion report creation and error log verification

Scored production Sheets data from T23; production Notion workspace targeted

Notion report page created in production workspace with all required sections; page URL written back to Sheets; automation error log is empty confirming zero unhandled exceptions across the full run

Notion page accessible via production URL; all five sections present; Sheets error log tab contains zero rows; execution log shows no unhandled exceptions

T25

Process owner review and go-live approval

HR Manager (process owner) reviews the Notion production report page created in T24; confirms accuracy of scores, narrative, and threshold flags; records approval with FullSpec team

Process owner confirms the report is accurate and the automation behaved as expected across the full cycle; go-live authorisation is recorded in the FullSpec confirmation log; Notion test page and Sheets test rows voided after approval

Process owner approval recorded in FullSpec confirmation log with name, date, and explicit go-live authorisation. Notion test page deleted. Sheets test rows cleared. FullSpec team notified to proceed to live deployment.

Once T25 is recorded as passed and the FullSpec confirmation log is updated, the automation is cleared for live deployment. The first production survey cycle will run against the full active employee list as configured in BambooHR. Any issues observed after go-live should be reported to support@gofullspec.com.

Test and QA PlanPage 4 of 4

Test and QA Plan

Test and QA Plan

01Testing approach

02Phase 1 — unit test cases

03Phase 2 — integration and edge case tests

04Phase 3 — end-to-end production test

More documents for this process